Electronic Journal of Differential Equations Vol. 1994(1994), No. 04, pp. 1-10. Published July 8, 1994. ISSN 1072-6691. URL: http://ejde.math.swt.edu or http://ejde.math.unt.edu ftp (login: ftp) 147.26.103.110 or 129.120.3.113 EXISTENCE RESULTS FOR NON-AUTONOMOUS ELLIPTIC BOUNDARY VALUE PROBLEMS V. Anuradha, S. Dickens, and R. Shivaji Abstract. We study solutions to the boundary value problems −∆u(x) = λf (x, u); u(x) + α(x) ∂u(x) = 0; ∂n x∈Ω x ∈ ∂Ω where λ > 0, Ω is a bounded region in RN ; N ≥ 1 with smooth boundary ∂Ω, α(x) ≥ 0, n is the outward unit normal, and f is a smooth function such that it has either sublinear or restricted linear growth in u at infinity, uniformly in x. We also consider f such that f (x, u)u ≤ 0 uniformly in x, when |u| is large. Without requiring any sign condition on f (x, 0), thus allowing for both positone as well as semipositone structure, we discuss the existence of at least three solutions for given λ ∈ (λn , λn+1 ) where λk is the k-th eigenvalue of −∆ subject to the above boundary conditions. In particular, one of the solutions we obtain has non-zero positive part, while another has non-zero negative part. We also discuss the existence of three solutions where one of them is positive, while another is negative, for λ near λ1 , and for λ large when f is sublinear. We use the method of sub-super solutions to establish our existence results. We further discuss non-existence results for λ small. 1. Introduction. We first consider the boundary value problem −∆u(x) = λf (x, u); u(x) + α(x) ∂u(x) = 0; ∂n x∈Ω x ∈ ∂Ω (1.1) (1.2) where ∆ is the Laplacian operator, Ω is a bounded region in RN ; N ≥ 1 with smooth boundary ∂Ω, α(x) ≥ 0, n is the unit outward normal, and f is a smooth function such that f (x, u) = σ uniformly in x, (1.3) lim u→∞ u Submitted: January 23, 1994. 1991 AMS SubjectClassification. 35J65. Partially supported by NSF Grants DMS - 8905936 and DMS - 9215027. This work was completed while R. Shivaji was a Visiting Research Scientist at Georgia Institute of Technology in 1993. c 1994 Southwest Texas State University and University of North Texas. 1 V. Anuradha, S. Dickens, and R. Shivaji 2 EJDE–1994/04 f (x, u) = β uniformly in x, u→−∞ u (1.4) ∂f (x, u) ≥ 0. ∂u (1.5) lim and Let λk , φk be the eigenvalues and corresponding eigenfunctions of the boundary value problem (1.6) −∆φk = λk φk ; x ∈ Ω φk (x) + α(x) ∂φk (x) = 0; ∂n x ∈ ∂Ω. (1.7) Let I = [a, b] ⊂ (λn , λn+1 ) and for λ ∈ I consider the unique solution Zλ of the boundary value problem −∆Zλ − λZλ = −1; Zλ (x) + α(x) x∈Ω ∂Zλ (x) = 0.; ∂n x ∈ ∂Ω (1.8) (1.9) Let µ1 = inf x∈Ω̄, λ∈I Zλ (x), µ2 = supx∈Ω̄, λ∈I Zλ (x) and µ = max{|µ1 |, µ2 }. Note that Zλ+ 6≡ 0 (see Appendix 1). Also, ∃ δ(Ω) > 0 such that if I ⊂ (λ1 , λ1 + δ) then Zλ > 0; x ∈ Ω (anti-maximum principle: see Clémént and Peletier in [1]). Now assume that ∃ m > 0 such that y + m > f (x, y) > y − m ∀ y ∈ [−bmµ, bmµ] and β, σ < 1 b||wα ||∞ (1.10) (1.11) where wα is the unique positive solution to −∆wα = 1; wα (x) + α(x) x∈Ω ∂wα (x) = 0; ∂n (1.12) x ∈ ∂Ω. (1.13) Then we prove: Theorem 1.1. Let I = [a, b] ⊂ (λn , λn+1 ) and (1.3)–(1.5), (1.10)–(1.11) hold. Then these exists at least three solutions to (1.1)–(1.2) for λ ∈ I. One of these solutions is a non-negative or sign changing solution, while another is a non-positive or sign changing solution. Remark 1.1. Note that unlike in the literature of “jumping nonlinearities” (see [2]– [3]), our results do not require (β, σ) to include a part of the spectrum of −∆, nor require, a part of the spectrum of −∆ to lie between f 0 (0) and f 0 (±∞) for autonomous nonlinearities as in [4] and the reference within. EJDE–1994/04 Non-autonomous EllipticBoundary Value Problems 3 Theorem 1.2. Let I = [a, b] ⊂ (λ1 , λ1 + δ) and (1.3)–(1.5), (1.10)–(1.11) hold. Then there exists at least three solutions to (1.1)–(1.2) for λ ∈ I, where one is a positive solution while another is a negative solution. Next we consider the particular case when σ = β = 0 (sublinear case). We further assume that there exists f1 (u) < f (x, u); ∀ x ∈ Ω̄, u ≥ 0 such that Z r1 0 f (s) ds > 0 (1.14) f1 (r1 ) = 0, f1 (r1 ) < 0, 0 for some r1 > 0, and that there exists f2 (u) > f (x, u); ∀ x ∈ Ω̄, u ≤ 0 such that Z 0 f2 (r2 ) = 0, f20 (r2 ) < 0, f (s) ds < 0 (1.15) r2 for some r2 < 0. Then we prove: Theorem 1.3. Assume (1.14)–(1.15) and let (1.3)–(1.5) hold with σ = β = 0. Then there exists at least three solutions to (1.1)–(1.2) for λ large, where one is a positive solution while another is a negative solution. Remark 1.2. We refer to [4] where a positive solution for λ ∈ I = [a, b] ⊂ (λ1 , λ1 +δ) was discussed in the case of autonomous, sublinear (σ = 0), and semipositone (f (0) < 0) problems with Dirichlet boundary conditions. In [5] it was assumed that there exists m > 0 such that f (y) ≥ y − m ∀y ∈ [0, bmµ] to obtain a positive solution for λ ∈ I. Further by constructing a function f1 (u) satisfying (1.14) a positive solution for λ large was discussed. See also [6] where the authors study the existence of a positive solution via degree theory arguments. Hence they allow non-autonomous problems with Robin boundary condition, but consider only the sublinear case, and study the existence of positive solutions for λ large with the assumption f (x, 0) ≤ 0 ∀x ∈ Ω̄. We also consider f such that ∃ r > 0 for which f (x, r) ≤ 0 while f (x, −r) ≥ 0 for every x ∈ Ω̄. (1.16) Then we prove: Theorem 1.4. Let I = [a, b] ⊂ (λn , λn+1 ) and (1.10), (1.16) hold. Assume r ≥ bmµ. Then there exists at least three solutions to (1.1)–(1.2) for λ ∈ I. One of these solutions is a non-negative or sign changing solution, while another is a non-positive or sign changing solution. Theorem 1.5. Let I = [a, b] ⊂ (λ1 , λ1 + δ) and (1.10), (1.16) hold. Assume r ≥ bmµ. Then there exists at least three solutions to (1.1)–(1.2) for λ ∈ I. One of these is a positive solution, while another is a negative solution. Theorem 1.6. Let (1.16) hold, and assume that (1.14) holds for all 0 ≤ u ≤ r with r1 ≤ r while (1.15) holds for all −r ≤ u ≤ 0 with r2 ≥ −r. Then there exists at least three solutions to (1.1)–(1.2) for λ large, where one is a positive solution while another is a negative solution. Remark 1.3. Consider the boundary value problem −∆u(x) = λf (||x||, u); x ∈ BN 4 V. Anuradha, S. Dickens, and R. Shivaji u(x) + α ∂u(x) = 0; ∂n EJDE–1994/04 x ∈ ∂BN where α ≥ 0 is a constant and BN is the unit ball in RN . Then under the corresponding hypothesis on f (||x||, u) instead of f (x, u), one can generate all the solutions obtained in Theorems 1.1–1.6 to be radial. This follows from the fact that all the sub and super solutions we will use in the proofs of these theorems will turn out to be radial. We next discuss non-existence results under the assumption that there exists γ > 0 such that ∂f ≤ γ. (1.17) ∂u We recall that λ1 is the principal eigenvalue and φ1 > 0 is a corresponding eigenfunction of −∆ subject to boundary Robin conditions (1.2). Then we prove: R Theorem 1.7. Assume Ω f (x, 0)φ1 (x) ≤ 0 and λ < λ1 /γ. If u(6≡ 0) is any solution of (1.1)–(1.2) then u− 6≡ 0 (i.e. there does not exist any solution u(6≡ 0) which is non-negative). R Theorem 1.8. Assume Ω f (x, 0)φ1 (x) ≥ 0 and λ < λ1 /γ. If u(6≡ 0) is any solution of (1.1)–(1.2) then u+ 6≡ 0 (i.e. there does not exist any solution u(6≡ 0) which is non-positive). Theorem 1.9. Assume f (x, 0) ≤ 0 and λ < λ1 /γ. If u is any solution of (1.1)– (1.2) then u ≤ 0. Theorem 1.10. Assume f (x, 0) ≥ 0 and λ < λ1 /γ. If u is any solution of (1.1)– (1.2) then u ≥ 0. We give detailed proofs of our results in section 2. For literature on autonomous semipositone problems with Dirichlet boundary conditions see [7], while for the positone case see [8] and the references cited in [8]. Our existence results are based on sub-super solutions. Namely, a super solution is defined as a smooth function φ such that −∆φ ≥ λf (x, φ); φ(x) + α(x) ∂φ(x) ≥ 0; ∂n x∈Ω x ∈ ∂Ω (1.18) (1.19) and a subsolution is a smooth function ψ that satisfies (1.18)–(1.19) with the inequalities reversed. If ψ ≤ φ, then it follows that (1.1)–(1.2) has a solution u such that ψ ≤ u ≤ φ (see [9]–[10]). Further if ψ1 is a subsolution, ψ2 is a strict subsolution, φ1 is a strict supersolution and φ2 is a supersolution such that ψ1 ≤ ψ2 ≤ φ2 , ψ1 ≤ φ1 ≤ φ2 , ψ2 (x0 ) > φ1 (x0 ) for some x0 ∈ Ω̄, then (1.1)–(1.2) has at least three distinct solutions u1 , u2 , u3 such that u1 ≤ u2 ≤ u3 . See [11] for this multiplicity result which was proved for Dirichlet boundary conditions. However, it is easy to see that this result holds for the Robin boundary conditions (1.2) as well. EJDE–1994/04 2. Non-autonomous EllipticBoundary Value Problems 5 Proofs of Theorems 1.1–1.10. Proof of Theorem 1.1. Let v1 (x) := bmZλ (x) and u2 (x) = −v1 (x). Then −∆v1 = bm(−∆Zλ ) = bm(λZλ − 1) ≤ bm(λZλ − λ/b)( since λ ≤ b) = λ[bmZλ − m] = λ[v1 − m] < λf (x, v1 ) by (1.10), and −∆u2 = ∆v1 ≥ λ[−v1 + m] = λ[u2 + m] > λf (x, u2 ) again by (1.10). Thus v1 is a strict subsolution while u2 is a strict supersolution. Note that v1+ 6≡ 0 and u− 2 6≡ 0. Now let u1 (x) := Jwα (x) where J > 0 is large enough so that f (x, J||wα ||∞ ) 1 ≥ b||wα ||∞ J||wα ||∞ (2.1) and u1 ≥ v1 , u1 ≥ u2 . (2.2) Here (2.1) is possible since σ < ||wα1||∞ b , and (2.2) is possible by the Hopf’s maximum principle. Then −∆u1 = J ≥ λf (x, J||wα ||∞ ); x ∈ Ω (by (2.1) since λ ≤ b) ˜ α (x) where J˜ > 0 ≥ λf (x, Jwα ) (since ∂f ≥ 0) = λf (x, u1 ). Next let v2 (x) = −Jw ∂u is large enough so that ˜ α ||∞ ) 1 f (x, −J||w ≥ ˜ α ||∞ b||wα ||∞ −J||w (2.3) and v2 ≤ v1 , v2 ≤ u2 . (2.4) 1 ||wα ||∞ b and (2.4) is possible by the Hopf’s ˜ ˜ α ||∞ ); x ∈ Ω (by (2.3) maximum principle. Then −∆v2 = −J ≤ λf (x, −J||w ∂f ˜ α ) (since since λ ≤ b) ≤ λf (x, −Jw ∂u ≥ 0) = λf (x, v2 ). Thus u1 is a supersolution while v2 is a subsolution such that v2 ≤ v1 ≤ u1 and v2 ≤ u2 ≤ u1 , where v1 is a strict subsolution, u2 is a strict supersolution with v2 ≤ 0, v1+ 6≡ 0, u− 2 6≡ 0 and u1 ≥ 0. Hence the result. Here again (2.3) is possible since β < Proof of Theorem 1.2. If λ ∈ (λ1 , λ1 + δ) then Zλ > 0; x ∈ Ω. Then v1 (x) > 0; x ∈ Ω while u2 < 0; x ∈ Ω and hence the result follows by the proof in Theorem 1.1. Proof of Theorem 1.3. Consider the autonomous Dirichlet problem −∆v = λf1 (v); v = 0; x∈Ω x ∈ ∂Ω. (2.5) (2.6) Then by (1.14) it follows that there exists λ̄1 > 0 such that for λ ≥ λ̄1 (2.5)λ (2.6) has a positive solution vλ (see [12]). Clearly since ∂v ∂n ≤ 0; x ∈ ∂Ω and f1 (S) < f (x, S); ∀x ∈ Ω̄, S ≥ 0 vλ is a strict subsolution to (1.1)–(1.2) for λ ≥ λ̄1 . Next consider (2.7) −∆u = λf2 (u); x ∈ Ω u = 0; x ∈ ∂Ω. (2.8) 6 V. Anuradha, S. Dickens, and R. Shivaji EJDE–1994/04 Then setting w = −u we see that w satisfies −∆w = λ[−f2 (−w)]; w = 0; x∈Ω x ∈ ∂Ω. (2.9) (2.10) R −r2 Let g(w) = −f2 (−w). Then g(−r2 ) = 0, g0 (−r2 ) = f20 (r2 ) < 0 and 0 g(s)ds R0 R −r2 = 0 −f2 (−s)ds = − γ2 f2 (s)ds > 0. Hence again by [12], there exists λ̄2 > 0 such that for λ ≥ λ¯2 , (2.9)–(2.10) has a positive solution wλ . Equivalently, (2.7)– λ (2.8) has a negative solution uλ = −wλ for λ ≥ λ̄2 . Also since ∂u ≥ 0; x ∈ ∂Ω ∂n and f2 (S) > f (x, S); ∀x ∈ Ω̄, S ≥ 0, uλ is a strict super solution to (1.1)–(1.2) for λ ≥ λ¯2 . Let λ̄ = max{λ¯1 , λ̄2 } and λ > λ̄ be fixed. Consider u1 (x) := Jwα (x) where J > 0 is large enough so that 1 f (x, J||wα ||∞ ) ≥ λ||wα ||∞ J||wα ||∞ (2.11) u1 ≥ vλ . (2.12) and Here (2.11) is possible since σ = 0 and (2.12) is possible by the Hopf’s maximum ∂f principle. Then −∆u1 = J ≥ λf (x, J||wα ||∞ ) ≥ λf (x, Jwα ) (since ∂u ≥ 0) = ˜ α (x) when J˜ > 0 is large enough so that λf (x, u1 ). Next consider v2 (x) := −Jw ˜ α ||∞ ) 1 f (x, −J||w ≥ ˜ α ||∞ λ||wα ||∞ −J||w (2.13) v2 ≤ uλ . (2.14) and Here again (2.13) is possible since β = 0 and (2.14) is possible by the Hopf’s ˜ α ||∞ ) ≤ λf (x1 , −Jw ˜ α ) (since maximum principle. Then −∆v2 = −J˜ ≤ λf (x, −J||w ∂f ∂u ≥ 0) = λf (x, v2 ). Thus u1 is a supersolution while v2 is a subsolution such that v2 ≤ uλ ≤ 0 ≤ vλ ≤ u1 , where uλ is a strict supersolution and vλ is a strict subsolution. Hence the result. Proof of Theorem 1.4. Let v1 (x) and u2 (x) be the strict sub and strict super solutions to (1.1)–(1.2) as in the proof of Theorem 1.1. Then since r ≥ bmµ, both v1 (x) and u2 (x) satisfy −r ≤ v1 (x) ≤ r, −r ≤ u2 (x) ≤ r. But by (1.16) u1 (x) = r and v2 (x) ≡ −r are super and sub solutions respectively to (1.1)–(1.2). Hence the result. Proof of Theorem 1.5. Let v1 (x) and u2 (x) be as in the proof of Theorem 1.4. But λ ∈ (λ1 , λ1 + δ). Hence v1 (x) ≥ 0 while u2 (x) ≤ 0. The rest of the proof is identical to the proof of Theorem 1.4. Proof of Theorem 1.6. Let uλ and vλ be respectively the strict super and strict subsolutions to (1.1)–(1.2) as in the proof of Theorem 1.3. But −r2 ≤ uλ ≤ 0 ≤ vλ ≤ r1 (see [12]) while u1 (x) ≡ r are v2 (x) = −r are super and subsolutions respectively to (1.1)–(1.2). Hence the result. EJDE–1994/04 Non-autonomous EllipticBoundary Value Problems 7 Proof of Theorem 1.7. Assume u ≥ 0, u 6≡ 0. Then −∆u = λf (x, u) = λ[f (x, u) − f (x, 0)] + λf (x, 0) ∂f = λ (x, η)u + λf (x, 0)( where 0 ≤ η ≤ u) ∂u ≤ λγu + λf (x, 0)( by (1.17)). Thus u satisfies −∆u − λγu ≤ λf (x, 0); x∈Ω ∂u(x) = 0; x ∈ ∂Ω. ∂n Multiplying (2.15) by φ1 and integrating we obtain Z Z Z −∆uφ1 dx − λγuφ1 dx ≤ λf (x, 0)φ1 dx ≤ 0. u(x) + α(x) Ω Ω (2.15) (2.16) Ω Applying Green’s second identity we obtain Z Z Z ∂u ∂φ1 {−φ1 uλ1 φ1 dx − λγuφ1 dx ≤ 0. +u } ds + ∂n ∂n ∂Ω Ω Ω But applying the boundary conditions we see that Z ∂u ∂φ1 {−φ1 +u } ds = 0. ∂n ∂n ∂Ω Thus we obtain Z uφ1 (λ1 − λγ) dx ≤ 0. Ω But u ≥ 0, φ1 > 0 in Ω and this is a contradiction if λ < λ1 /γ. Hence the result. Proof of Theorem 1.8. Assume u ≤ 0, u 6≡ 0. Proceeding as in the proof of Theorem 1.7 we obtain ∂f (x, z)u + λf (x, 0) ∂u ≥ λγu + λf (x, 0)( by (1.17)). −∆u = λ Thus u satisfies −∆u − λγu ≥ λf (x, 0); x∈Ω (2.17) ∂u(x) = 0; x ∈ ∂Ω. (2.18) u(x) + α(x) ∂n R Multiplying (2.17) by φ1 , using Ω f (x, 0)φ1 dx ≥ 0, and proceeding as in the proof of Theorem 1.7 we obtain Z uφ1 (λ1 − λγ) dx ≥ 0. Ω 8 V. Anuradha, S. Dickens, and R. Shivaji EJDE–1994/04 But u ≤ 0 while φ1 > 0 for x ∈ Ω, thus this is a contradiction if λ < λ1 /γ. Hence the result. Proof of Theorem 1.9. Suppose u > 0 somewhere in Ω̄. Then ∃ some Ω1 ⊆ Ω such that u > 0 in Ω, and either ∂u(x) u(x) = 0 or u(x) + α(x) = 0; x ∈ ∂Ω. (2.19) ∂n Let λ̃1 be the principal eigenvalue and φ̃1 > 0 be a corresponding eigenfunction of −∆ on the region Ω1 subject toRRobin conditions (1.2) on ∂Ω1 . Note that λ̃1 ≥ λ1 . Now since f (x, 0) ≤ 0 we have Ω f (x, 0)φ̃1 dx ≤ 0. Thus following the steps in the proof of Theorem 1.7 we obtain Z Z −∆uφ̃1 dx − λγuφ̃1 dx ≤ 0 Ω1 and then Ω1 Z ∂u ∂ φ̃1 [−φ̃1 +u ] ds + ∂n ∂n ∂Ω1 Z [λ̃1 − λγ]uφ̃1 dx ≤ 0. Ω1 But if u = 0 on ∂Ω1 , then ∂u ≤ 0 on ∂Ω1 , while if u + α ∂n = 0 on ∂Ω1 , since φ̃1 φ̃1 + α ∂∂n = 0, we have φ̃1 + u ∂∂n = 0 on ∂Ω1 . Thus in any case (see (2.19)) ∂u ∂n ∂u −φ̃1 ∂n φ̃1 ∂u −φ̃1 ∂n + u ∂∂n ≥ 0. Hence Z [λ̃1 − λγ]uφ̃1 dx ≤ 0 Ω1 which is a contradiction since u > 0, φ̃1 > 0 for x ∈ Ω1 while λ < λ1 /γ ≤ λ˜1 /γ. Hence the result. Proof of Theorem 1.10. Suppose u < 0 somewhere in Ω̄. Then ∃ some Ω1 ⊆ Ω such that u < 0 in Ω1 and either ∂u(x) u(x) = 0 or u(x) + α(x) (2.20) = 0; x ∈ ∂Ω1 ∂n RLet λ̃1 and φ̃1 > 0 be as in the proof of Theorem 1.9. Now f (x, 0) ≥ 0 and hence f (x, 0)φ̃1 dx ≥ 0. Thus following the steps in the proof of Theorem 1.8 we obtain Ω1 Z Z −∆uφ̃1 dx − λγuφ̃1 dx ≥ 0, Ω1 and then Z Ω1 ∂u ∂ φ̃1 [−φ̃1 +u ] ds + ∂n ∂n ∂Ω But if u = 0 on ∂Ω1 , then ∂u ∂n Z [λ̃1 − λγ]uφ̃1 dx ≥ 0. Ω1 ∂u ≥ 0 on ∂Ω1 , while if u + α ∂n = 0 on ∂Ω1 , since φ̃1 φ̃1 ∂u φ̃1 + α ∂∂n = 0 on ∂Ω, we have −φ̃1 ∂n + u ∂∂n = 0 on ∂Ω1 . Thus in any case (see φ̃1 ∂u (2.20)) −φ̃1 ∂n + u ∂∂n ≤ 0. Hence Z [λ̃1 − λγ]uφ̃1 dx ≥ 0 Ω1 which is a contradiction since u < 0, φ̃1 > 0 for x ∈ Ω, while λ < λ1 /γ ≤ λ̃1 /γ. Hence the result. EJDE–1994/04 Non-autonomous EllipticBoundary Value Problems 9 Appendix 1. Let λ ∈ (λn , λn+1 ) and consider the unique solution to the boundary value problem −∆Zλ − λZλ = −1; x ∈ Ω Zλ (x) + α(x) ∂Zλ (x) = 0; ∂n x ∈ ∂Ω. Let φ1 > 0 satisfy (1.6)–(1.7) for k = 1. Then Z Z Z −∆Zλ φ1 dx − Ω −φ1 dx λZλ φ1 dx = Ω Ω which implies Z ∂Zλ ∂φ1 {−φ1 + Zλ } ds + ∂n ∂n ∂Ω Z Z Zλ λ1 φ1 dx − Ω Z −φ1 dx . λZλ φ1 dx = Ω Ω But Z ∂Zλ ∂φ1 {−φ1 + Zλ } ds = ∂n ∂n ∂Ω Thus Z {α ∂Ω ∂φ1 ∂Zλ ∂Zλ ∂φ1 −α } ds = 0 . ∂n ∂n ∂n ∂n Z Z (λ1 − λ)Zλ φ1 dx = Ω −φ1 dx. Ω But λ > λ1 and φ1 > 0 for x ∈ Ω. Hence Zλ+ 6≡ 0. References 1. Ph. Clément and L. A. Peletier, An anti-maximum principle for second order elliptic operators, J. Diff. Eqns. 34 (1979), 218–229. 2. A. C. Lazer and P. J. McKenna, Large-Amplitude periodic oscillations in suspension bridges: some new connections with nonlinear analysis, SIAM Rev. 32(4) (1990), 537–578. 3. A. Castro and S. Gadam, The Lazer-McKenna conjecture for radial solutions in the RN ball, Electronic J. Diff. Eqns. 7 (1993), 1–6. 4. A. Castro and J. Cossio, Multiple solutions for a nonlinear Dirichlet problem, SIAM Journal of Math. Anal. (to appear). 5. A. Castro, J. B. Garner and R. Shivaji, Existence results for classes of sublinear semipositone problems, Results in Mathematics 23 (1993), 214–220. 6. W. Allegretto and P. Nistri, Existence and stability for non-positone elliptic problems, J. Nonlinear Analysis TMA (to appear). 7. A. Castro and R. Shivaji, Semipositone problems, Semigroups of Linear and Nonlinear Operations and applications (1993), 109–119. 8. K. J. Brown, M. M. A. Ibrahim and R. Shivaji, S-shaped bifurcation curves, J. Nonlinear Analysis, TMA 5(5) (1981), 475–486. 9. H. Amann, Fixed point equations and nonlinear eigenvalue problems in ordered Bannach spaces, SIAM Rev. 18 (1976), 620–709. 10. D. H. Sattinger, Monotone methods in nonlinear elliptic and parabolic boundary value problems, Indiana University Mathematics Journal 21(11) (1972), 979–1000. 11. R. Shivaji, A remark on the existence of three solutions via sub-super solutions, Nonlinear Analysis and Applications (1987), 561–566. 10 V. Anuradha, S. Dickens, and R. Shivaji EJDE–1994/04 12. P. Clement and G. Sweers, Existence and multiplicity results for a semilinear elliptic eigenvalue problem, Ann. Scuola Norm. Sup. Pisa. Cl. Sci. (4)14 (1987), 97–121. V. Anuradha, Department of Mathematics and Statistics, University of Arkansas at Little Rock, Little Rock, AR 72204-1099 S. Dickens, Department of Mathematics and Statistics, Mississippi State University, Mississippi State, MS 39762 R. Shivaji, Department of Mathematics and Statistics, Mississippi State University, Mississippi State, MS 39762. E-mail rs1@ra.msstate.edu